Vanadium(I) chloride and lithium vanadium(I) dihydride as epimetallating reagents for unsaturated organic substrates: Constitution and mode of reaction

Article abstract of DOI:10.1002/ejoc.200800104

Subvalent vanadium(I) salts, of empirical formulas, VCl, vanadium(I) chloride and LiVH₂, lithium vanadium(I) dihydride, can be conveniently prepared in THF solution, starting at -78 °C, by treating either VCl₃ or VCl₄ with an appropriate number of equivalents of nBuLi. As judged by the stability of solutions or solid samples of LiVH₂, the preparation of LiVH₂ from VCl₄ is the preferred method. Individual physical characterization of solid samples of VCl or of LiVH₂, admixed with their LiCl by-product, was carried out after removal of all volatiles in vacuo and by the following measurements: 1) gasometric protolysis with glacial acetic acid and measurement of the H ₂ evolved in the oxidation of V^I to V^II; 2) infrared spectroscopic search for V-H bands; and 3) examination for unpaired electrons by EPR activity. Such measurements applied to VCl lend strong support for a V^I oxidation state but only probable evidence for paramagnetism and for the association of VCl units. Similar measurements applied to LiVH ₂ give unambiguous gasometric and IR evidence favoring the LiVH ₂ stoichiometry and the biradical nature of the VH₂ anion with a linear array of H-V-H atoms. Chemical characterization of both VCl and LiVH₂ toward individual organic substrates, such as olefins, ketones, epoxides and organic halides, yielded convincing evidence that organic radical mechanisms are involved, both for the proven biradical, LiVH₂, as well as for the diamagnetic VCl. Finally, the question of why LiVH₂ prepared from VCl₄ is more stable than the LiVH₂ obtained from VCl₃ is addressed in terms of the actual coordination sphere of the VH₂ anion in THF solution and in the solid state. Preliminary studies comparing the reactivities of LiVH₂ and LiCrH₂ toward organic substrates indicate that LiVH₂ is the distinctly more moderate and usefully selective reductant. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200800104

FULL PAPER
DOI: 10.1002/ejoc.200800104
Vanadium(I) Chloride and Lithium Vanadium(I) Dihydride as Epimetallating
Reagents for Unsaturated Organic Substrates: Constitution and Mode of
Reaction[‡]
John J. Eisch,*[a] Paul O. Fregene,[a] and David C. Doetschman[a,b]
Dedicated to Professor Dr. Dr. h. c. mult. Wolfgang A. Herrmann on the occasion of his 60th birthday
–
Keywords: Epimetallation / EPR spectroscopy of VH2 / Vanadium(I) chloride / Lithium vanadium(I) dihydride / Organic
radical intermediates
Subvalent vanadium(I) salts, of empirical formulas, VCl, va-
nadium(I) chloride and LiVH2, lithium vanadium(I) dihy-
dride, can be conveniently prepared in THF solution, starting
at –78 °C, by treating either VCl3 or VCl4 with an appropriate
number of equivalents of nBuLi. As judged by the stability
of solutions or solid samples of LiVH2, the preparation of
LiVH2 from VCl4 is the preferred method. Individual physical
characterization of solid samples of VCl or of LiVH2, admixed
with their LiCl by-product, was carried out after removal of
all volatiles in vacuo and by the following measurements: 1)
gasometric protolysis with glacial acetic acid and measure-
ment of the H2 evolved in the oxidation of VI to VII; 2) infra-
red spectroscopic search for V–H bands; and 3) examination
for unpaired electrons by EPR activity. Such measurements
applied to VCl lend strong support for a VI oxidation state
but only probable evidence for paramagnetism and for the
association of VCl units. Similar measurements applied to
LiVH2 give unambiguous gasometric and IR evidence favor-
ing the LiVH2 stoichiometry and the biradical nature of the
VH2 anion with a linear array of H–V–H atoms. Chemical
characterization of both VCl and LiVH2 toward individual or-
ganic substrates, such as olefins, ketones, epoxides and or-
ganic halides, yielded convincing evidence that organic radi-
cal mechanisms are involved, both for the proven biradical,
LiVH2, as well as for the diamagnetic VCl. Finally, the ques-
tion of why LiVH2 prepared from VCl4 is more stable than
the LiVH2 obtained from VCl3 is addressed in terms of the
actual coordination sphere of the VH2 anion in THF solution
and in the solid state. Preliminary studies comparing the re-
activities of LiVH2 and LiCrH2 toward organic substrates
indicate that LiVH2 is the distinctly more moderate and use-
fully selective reductant.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
Mg, Al or Zn, cf. Equation (1). This transformation, which
has been explored by the McMurry group[5] and a host of
other researchers,[6] is a most valuable method for the as-
sembly of a great variety of open-chain and ring-carbon
skeletons. Unfortunately, the great number of recipes of-
fered for generating the active subvalent titanium reagent
have provided little insight into the actual oxidation state
involved, Ti0 and/or TiII, nor into the role that any ligands
on Ti or an excess of main-group metal reductant might
play in the reaction mechanism of C–C bond formation.[7]
Introduction
Subvalent Early Transition Metal Reductants in Organic
Synthesis
The great utility and versatility of subvalent early transi-
tion metal salts as reductants in organic chemistry became
evident in 1973, when the research groups of Mukaiyama,[2]
Tyrlik[3] and McMurry[4] reported independently on the
novel reductive dimerization of organic carbonyl derivatives
by various combinations of titanium(III) or titanium(IV)
salts with a main-group metal reductant, such as LiAlH4,
[‡] Organic Chemistry of Subvalent Transition Metal Complexes,
43. Part 42: Ref.[1]
[a] State University of New York at Binghamton, Department of
Chemistry,
Therefore, over the last 15 years our group has endeav-
ored to develop preparative routes to defined oxidation
states of subvalent titanium and other early transition met-
als, which reagents would be soluble and sufficiently stable
in various ethers and would be free of any residual main-
group reductant. Our first target was TiCl2,[7–9] where we
P. O. Box 6000, Binghamton, NY 13902-6000, USA
Fax: +1-607-777-4865
E-mail: jjeisch@binghamton.edu
[b] Regional Center for Pulsed EPR & Photochemical Studies,
State University of New York at Binghamton, Department of
Chemistry,
P. O. Box 6000, Binghamton, NY 13902-6000, USA
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2825

J. J. Eisch, P. O. Fregene, D. C. Doetschman
FULL PAPER
were pleased to discover that it could be prepared essen- Oxidative Additions of Neutral and Anionic Transition-
tially quantitatively by the process given in Scheme 1. The Metal Complexes to Unsaturated Covalent Bonds
TiCl4 (1), dissolved in THF at –78 °C, was treated at –78 °C
The great potential of such transition reductants in or-
with two equivalents of BuLi in hexane to generate the tan-
ganic synthesis is their oxidative addition to π-bonded
colored compound 2. By raising the temperature thereafter
C=C, CϵC, C=O or C=N functional groups. The first re-
to 25 °C, the butyl groups were lost, as butane and 1-but-
cognition of the vital role of such an addition was the dis-
ene, and titanium(II) chloride (3) was isolated as a black
covery of the high-yielding synthesis of cyclopropanols (8)
solid. Although 3 can be isolated free of LiCl as the com-
by Kulinkovich and co-workers from the interaction of es-
plex, TiCl2·2THF (3a), there is evidence that in the original
ters with ethyl Grignard reagents and titanium(IV) isopro-
solution 3 may also form complexes with 1-butene or with
poxide.[14] In rationalizing this surprising result it was sug-
LiCl (cf. infra). Thus TiCl2 should be considered as the em-
gested that the three-membered intermediate 7 was in-
pirical formula of 3 and not representative of the molecular
volved.[15] Intermediate 7 can be described by two extreme
formula, the state of solvation or the actual coordinated
resonance structures: 7a as an η2-π complex with little elec-
number of ligands.
tron-transfer to the organic ligand and 7b as a metallacyclo-
propane with much electron-transfer from the metal
(Scheme 3). Strong support for the intermediacy of 7 in the
Kulinkovich reaction was brought forward from our group
by preparing 7 separately and then showing that the two
Scheme 1.
apparent Ti–C bonds of 7b can be detected by stepwise in-
sertions of benzonitrile and then of CO2 (Scheme 4).[16]
The two-step process depicted in Scheme 1, known as
alkylative reduction, has subsequently been applied to pre-
Products 10 and 11 could only arise via intermediates 7b
and 9.
paring subvalent reductants having empirical formulas of
Ti(OiPr)2 from Ti(OiPr)4,[7,10] ZrCl2 from ZrCl4,[7,11] HfCl2
from HfCl4,[7,11] and CrCl (4) from CrCl3.[12]
Adding significantly to potential reductants obtainable
through alkylative reduction was the surprising finding that
anionic hydride complexes could be generated by an excess
of BuLi over the replaceable chlorides on the transition
metal (Scheme 2). Chromium(III) chloride (5), for example,
can give rise in high yield to lithium chromium(I) dihydride
(6).[13] Here again, as with the empirical formula of the fore-
going neutral subvalent complexes, it can be assumed that
the coordination sphere of 6 could be expanded by sol-
vation with THF molecules at the Li and Cr centers and/or
by coordination of chloride anions or of η2-bonded 1-but-
ene units at the chromium atom.
Scheme 3.
Such oxidative additions to π-bonds of C=C, CϵC, C=O
and C=N linkages, leading to three-membered metall-
ocyclic intermediates, as are involved in the reactions de-
Scheme 2.
picted in Equations (2)–(4), and the oxidative additions to
Scheme 4.
2826
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 2825–2835

Epimetallating Reagents for Unsaturated Organic Substrates
sigma-covalent bonds leading to bond cleavage and C–C
The remarkable properties of these chromium reagents
bond coupling, as illustrated in Equations (5)–(7), have have therefore motivated us to attempt the preparation of
been termed by our group as epimetallations (Greek: liter- the analogous derivatives of vanadium, namely VCl (12)
ally metallations upon a bond, by analogy with epoxidation and LiVH2 (13). Such reagents would lack one proton and
of an olefin).[17]
one electron in the metal atom, compared with their chro-
mium counterparts. There have been several reports of tre-
ating vanadium salts with RLi or RMgX reagents and uti-
lizing the unexamined reagent solution for applications in
organic synthesis but no systematic study of such vanadium
reactants has yet appeared.[21] Therefore, it was of great
interest to learn what differences in properties such vana-
dium reagents would exhibit in their changed electronic en-
vironment.
Results and Discussion
Preparation and Gasometric Analysis of Vanadium(I)
Chloride (12)
General Observation
A most reliable and convenient assay for the subvalent
oxidation state of an early transition metal, when the
amount of metal is already known to be present from the
initial weight of salt introduced, and for the presence of M–
H bonds, is the protolysis of the solid metal salt under an
inert atmosphere. Vanadium compounds of lower than VII
oxidation state are known to release one equivalent of H
for each unit of oxidation and of course each M–H bond
would release one equivalent of H2. Thus V0 would release
two H and VI one H. This type of gasometric analysis has
been applied to TiII,[8] ZrII[19] and CrI[12,13] complexes.
Samples of VCl (12), admixed with LiCl by-product,
were prepared and isolated from THF solution as a black
solid, either by treating purple VCl3 with two equivalents of
n-butyllithium or by reacting red-brown VCl4 with
three equivalents of n-butyllithium at –78 °C and then
warming to room temp., cf. Equation (8).
The selective conversion to specific reduction products
by these Group 4 reductants has spurred our interest in
studying the reactivity of Groups 5 and 6 neutral and an-
ionic reducing agents.
After prolonged pumping under high vacuum to remove
volatiles, each sample was subjected to gasometric analysis
of gas liberated by treatment with acetic acid. Three typical
samples gave gas volumes of 0.67, 0.63 and 0.55 mol at STP,
which gas by mass spectral analysis consisted of ca. 90% of
H2 and various small amounts of THF, butane and a but-
ene, as judged by minor mass peaks, cf. Equation (9). Such
an amount of H2 (0.56 mol of H2 as an average) is in rea-
sonable agreement with the 0.50 mol of H2 at STP expected
from Equation (9).[22]
Goal of the Present Investigation
In our aforementioned studies of CrCl (4)[12] and LiCrH2
(6),[13] we found that these reagents have proved to be the
most powerful and chemically versatile of the early transi-
tion metal reductants thus far examined. Chromium(I)
chloride (4) is capable of the efficient reductive coupling of
carbonyls to pinacols, of α-bromo ketones to 1,4-diketones
and of benzylic halides variously to bibenzyls or, with de-
halogenation, to diarylethenes or to diarylethynes.[12] Lith-
ium chromium(I) dihydride (13) shows a high reactivity in
2 VCl + 2 HOAc Ǟ 2 VCl(OAc) + H2Ȇ (9)
That the butane or butene in the gas evolved could well
cleaving C–O, C–S, C–N and C–X sigma-bonds and in have arisen from 1-butene remaining in the solid VCl as
oligomerizing acetylenes and olefins.[13]
epimetallated product 14 was demonstrated as follows.
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2827

J. J. Eisch, P. O. Fregene, D. C. Doetschman
FULL PAPER
When a solution of 12 in THF was treated with CO2 at
–78 °C and then hydrolyzed, an 11% yield of pentanoic acid
was obtained. This supports the existence of 14 in solution
and its trapping by carbonation[10], cf. Equation (10).
The source of the greater instability of the LiVH2 ob-
tained from the VCl3 starting material may stem from its
impurities (3%) or from the smaller proportion of the LiCl
by-product (LiVH2/4LiCl), which LiCl may help to stabilize
the LiVH2 by chloride coordination. Therefore, for the
more sensitive reactions of LiVH2, its preparation from
VCl4 is to be preferred.[23]
Preparation and Gasometric Analysis of Lithium
Vanadium(I) Dihydride (13)
Infrared Examination of Vanadium(I) Chloride (12) and
Lithium Vanadium(I) Dihydride(13)
Samples of LiVH2 (13), admixed with LiCl by-product,
were prepared from THF solution as a black solid, either
by treating purple VCl3 (97%) with four equivalents of n-
butyllithium or by allowing red-brown VCl4 (99% mini-
mum) to react with five equivalents of n-butyllithium at
–78 °C and warming to room temp., cf. Equation (11). The
LiVH2 reagents of a given formal concentration prepared
from either source had comparable reducing actions
towards ketones as to kinds and proportions of reduction
products but there are noteworthy differences. For example,
LiVH2 prepared from VCl4 yielded 75% of the principal
reduction products, 9-fluorenol and fluorene, in a ratio of
50:50, while LiVH2 from VCl3 gave a total yield of these
two products of 70% but in a ratio of 81:19. Thus the
LiVH2 reagent from VCl4 is somewhat a more potent re-
ductant.
Examination of a solid sample of the VCl/LiCl mixture
(12) as a mineral oil mull by infrared spectroscopy showed
the absence of any sharp absorptions, other than those of
the mineral oil, in the 1500–1650 cm–1 region and thus con-
firmed the absence of any V–H bonds. On the other hand,
solid samples of the LiVH2/LiCl mixture (13) generated
from VCl4 showed an absorption of medium intensity at
1625 cm–1 [24] and one of weak intensity at 1511 cm–1. By
comparison, the paramagnetic Cp2V–H exhibits a medium
band at 1625 cm–1 [25] and the paramagnetic [CpV(µ-H)-
dmp]2 a band at 1570 cm–1.[24] Absorptions of this LiVH2
at these two wavenumbers indicates the presence of a ter-
minal V–H bond and a bridging V–H bond.
The foregoing gasometric and infrared spectroscopic
data, therefore, support the conclusion that the solid
LiVH2/LiCl mixture obtained from VCl4 has chemical, pro-
tolytic and infrared spectral properties consistent with the
empirical formula of LiVH2.
Electron Paramagnetic Resonance Spectroscopy of
Vanadium(I) Chloride (12) and Lithium Vanadium(I)
Dihydride (13)
The measured differences in the properties of LiVH2
samples generated from VCl3 or from VCl4 became much
more pronounced when the LiVH2 samples were allowed to
stand in THF at room temp. or especially when the solvent
was completely removed to produce a solid mixture of sup-
posedly LiVH2 and LiCl. Acetolysis of such samples of
LiVH2 should theoretically evolve 2.5 mol of H2 per vana-
dium atom at STP per mol of 13, cf. Equation (12).[22]
General Observations
Solutions of the VCl/LiCl mixture in THF at 30 K gave
a broad but undefined absorption, indicating the presence
of some unknown paramagnetic component. Possibly this
may be monomeric VCl, formed from the dissociation of
a diamagnetic oligomer, (VCl)n but further studies will be
required to clarify this possibility.[26]
The CW-EPR spectrum of the flash-frozen solution of
lithium vanadium(I) dihydride (13) prepared from VCl3 in
a 1:1 heptane/toluene solution at 30 K, is the most revealing
of the biradical paramagnetism and bonding in the dihydri-
In fact, samples of LiVH2 (13) prepared from VCl4 did dovanadium(I) anion in 13. The CW-EPR spectrum of the
generate 2.5Ϯ0.4 mol of H2 at STP. In sharp contrast, how- flash-frozen 0.09  sample in 1:1 toluene/heptane solvent at
ever, acetolysis of the LiVH2 generated from VCl3 provided 30 K is shown in Figure 1. At temperatures of 140 K and
between 1.0–2.0 mol of H2 per vanadium atom at STP, val- above, the spectrum collapses into a single central line with
ues consistent with extensive decomposition of 13 into LiH, little or no hyperfine structure. A weak, but prominent, cen-
cf. Equations (13) and (14).
tral peak is also evident in Figure 1 and is present in all but
2828
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 2825–2835

Epimetallating Reagents for Unsaturated Organic Substrates
the least concentrated (0.0108 ) sample. The intensity of
The recorded fields of the transitions were fitted by least-
this peak was much weaker in samples of 13 prepared from squares with the parameters in the following spin Hamilto-
VCl4. The spectrum with the hyperfine structure in Fig- nian, cf. Equation (15).
–
ure 1, which will be shown to belong to the VH2 , persists
H = H·g·S + D [Sz2 – (1/3)S(S + 1)] + E(Sx2 ) + Σi = 1,2 S·AiH·Ii,
– y2
in the least concentrated sample. This central peak observed
(15)
at low temperatures is an impurity evidently associated with
a side reaction.
In Equation (15) gʈ is the zz principal value of g, gЌ is
H
the xx and yy principal value, and likewise AʈH and AЌ
are the corresponding, identical, principal values of
AiH.[27,28] A satisfactory fit to the data was obtained by as-
suming that the two proton hyperfine coupling constants
for the two protons, i = 1,2, are the same, as was indicated
by the evident 1:2:1 patterns in the spectrum. The electron
spin multiplicity is triplet, S = 1, and the proton spins have
Ii = 1/2. The results of the fit, together with the results of
the uncertainty analysis, are given in Table 1. The fact that
the value of E = 0 within experimental error confirms the
assumption of axial symmetry for the g and A matrixes.
Table 1. Spin Hamiltonian parameters for the ground triplet state
of VH2 , determined from the EPR spectrum of a 0.14  prepara-
tion at 30 K.
–
Figure 1. The CW-EPR spectrum of the flash-frozen 0.09  sample
of 13 in 1:1 toluene/heptane solvent at 30 K The 1:2:1 proton hy-
perfine multiplets and principal axis assignments are indicated with
lines.
Parameter
Value
Uncertainty
gʈ
gЌ
2.011
2.005
Ϯ377
Ϯ3
Ϯ144
Ϯ71
0.003
0.003
5
3
6
6
In order to estimate the proportion of this paramagnetic
impurity, the aforementioned sample of 13 was compared
with an EPR intensity standard to make a “spin count”.
D/geβe (Gauss)
E/geβe (Gauss)
AʈH/geβe (Gauss)
The double integral of the triplet portion of the spectrum AЌH/geβe (Gauss)
corresponded to about 0.20  in the triplet species. These
concentrations we believe to be equal within the precision
of the EPR experiment. The narrow central peak near the
center of the triplet spectrum is clearly an impurity, which
was most evident in EPR spectra of LiVH2 samples pre-
pared from VCl3 (cf. supra). Double integration of its inten-
sity shows it to be present at a concentration of no more
than 10% of the concentration of the triplet species.
One can break the A values down into an isotropic part,
AisoH/geβe = Ϯ 95.3 Gauss, and an anisotropic part whose
H
AanisoHʈ/geβe = Ϯ 48.7 Gauss and whose Aaniso Ќ/geβe =
–/+ 34.3 Gauss. (We reject the unlikely result, AisoH/geβe =
H
Ϯ 0.7 Gauss, from AʈH and AЌ being of opposite sign.)
These values will be discussed in the next section in terms
–
of the shape of VH2 and its electron spin distribution and
It is interesting to note the absence of any resolvable va-
nadium hyperfine structure. It is possible, however, that va-
nadium hyperfine interactions from this anion are distorted
by the solvent molecules and thus contribute to the surpris-
ingly intense, featureless background underlying the rela-
tively narrow hyperfine features. Alternatively, the broad
background may be due to a paramagnetic impurity.
electronic structure.
The triplet ground electronic state of the paramagnetic
system being observed by EPR is of axial symmetry, based
on the parameters given in Table 1. Principal g values and
principal A values of parallel axial symmetry fit the spec-
trum well. The fact that the spectrum is also fit well with E
= 0 within experimental error also points, as do the axial g
values, to an unpaired electron spin distribution of axial
symmetry. The Hamiltonian, based on the presence of two
I = 1/2 nuclei in parallel axial environments is consistent
with a linear molecular system with two equivalent H atoms,
as depicted in 15. It remains probable but unknown that
THF is coordinated both with the cationic Li and anionic
VH2.[29]
Analysis of the EPR Data for Lithium Vanadium(I)
Dihydride (13)
The 30 K pulsed EPR spectrum, which had somewhat
lesser resolution than the 30 K CW-EPR spectrum, con-
firmed the presence of peaks at the inflections indicated in
Figure 1. What appeared to be 1:2:1 hyperfine multiplets at
the low field z (2897.4 G, 3063.5 G, and 3188.2 G), low field
Let us next consider further the implications of the val-
ues of the parameters in Table 1 under the hypothesis that
the spectrum is from the symmetric, linear VH2 anion.
x
(3182.2 G, 3249.4 G, and 3307.5 G), high field x
(3540.1 G, 3604.7 G, and 3698.6 G), and high field z
(3663.8 G, 3793.0 G, and 3944.9 G) transitions, were care-
fully measured from the digital records of the spectra. The
y axis features could not be distinguished.
The isotropic part of the AiH for the two protons was
found to be AisoH/geβe = Ϯ 95.3 Gauss. On the basis on the
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2829

J. J. Eisch, P. O. Fregene, D. C. Doetschman
FULL PAPER
H
known Aiso = 508 Gauss of the H atom the value of the with the corresponding chromium reagents 4 and 6. The
–
H
VH2 Aiso implies that 19% of the triplet electron spin reactions with olefins, ketones, epoxides and halides have
– [27,28]
resides in each H 1s orbital of VH2 .
proved to be especially revealing and are given here.
The value of the observed D parameter is too small for
the electron spin dipolar interaction in a triplet molecule in
which the molecular orbitals of the two unpaired electrons
involve different atomic orbitals of the same atom.[30] The
relatively small g shifts from the free electron g value,
2.0023, preclude large spin-orbital contributions. Therefore,
in the small VH2 molecular ion this implies that one un-
paired electron is situated around the H atoms and the
other resides near the V atom, nearly exclusively.
Let us assume that the bonding of the H atoms with the
V atom is of the σ type between the H 1s orbitals and the
3dz2 orbital of the V atom. Simple molecular orbital consid-
erations would predict a bonding σ orbital with no nodes
perpendicular to the molecular axis, an essentially non-
bonding sigma type orbital (n) with a node through the V
nucleus, and an antibonding σ* orbital with two nodes cut-
ting the V–H bonds. Two of the eight electrons of this sys-
tem fill the bonding σ orbital, leaving six electrons to fill
the remaining orbitals of the molecular ion. One would ex-
pect the dxz, dyz, dx2–y2, and n orbitals, all of essentially non-
bonding character, to be moderately close in energy.
Olefins
Neither olefins nor acetylenes undergo epimetallation
readily but do undergo cis-trans isomerization efficiently.
cis-Stilbene (16) is isomerized by catalytic amounts of either
12 or 13 to more stable trans-stilbene (17). For the reaction
with 13 one could propose a radical addition-elimination
(via 18) as the operative pathway, cf. Equation (16). As a
reasonable extension, it is likely that with VCl (12), the pro-
posed diamagnetic oligomer (VCl)n could produce VCl radi-
cals by dissociation. The alternative suggestion by a referee
that such isomerization could be due to the Lewis acidity
of VCl is most unlikely: in THF any monomeric VCl would
be completely complexed with THF solvent and thus be
rendered only weakly electrophilic.
–
We suggest that the only configuration that satisfies
the need for one unpaired electron essentially entirely
on the H atoms and that maintains axial symmetry is
σ2 dxz dyz d1
n1. This idealization places 50% of the
x2–y2
unpaired electron density in the V dx2–y2 orbital and 50%
in the mainly H atom n σ-type orbital. Thus, were there no
contribution to the n orbital and were there only 1s H char-
acter in the n orbital, then 25% of the unpaired electron
density would reside in each H atom 1s orbital. Given that
these absolutes will not fully obtain, the observed 19% un-
paired spin density in the H 1s orbitals is further evidence
Another reaction consistent with the operation of free
radicals is the observed polymerization of styrene (19) into
atactic polymer at 25 °C, again by catalytic amounts of 12
or 13. Here initiation by 12 or 13 to form a biradical inter-
mediate similar to 18 would be a reasonable pathway, cf.
Equation (17).
–
for the essential truth of this simple description of the VH2
molecular anion. Clearly the small, symmetric VH2– molec-
ular anion would be an attractive candidate in which to
test these hypotheses with quantum mechanical calculations
H
and to compare calculated values of Aaniso
and
ʈ
H
Aaniso
with the corresponding experimental values.
Ќ
Chemical Characterization of Vanadium(I) Chloride (12)
and Lithium Vanadium(I) Dihydride (13)
Ketones
With reliable methods for the preparation and analysis
of THF solutions of VCl (12) and LiVH2 (13) reagents now
The behavior of the hindered ketone, 9-fluorenone (20)
in hand, we have launched a study of their applications in toward 12 and toward 13 offers an instructive comparison:
organic synthesis. The need for such a study becomes ap- two equivalents of VCl causes a 91% conversion of 20 into
parent when one notes that a review of vanadium in mod- reduced dimers at 25 °C, cf. Equation (18). When con-
ern organic synthesis published in 1997 has only 18 ci- ducted at reflux, 1.25 equivalents of 12 produces 21 in 99%
tations out of 139 references possibly involving subvalent yield. This dimerization could readily be ascribed to dimer-
vanadium.[21] We hope to publish our complete results in ization via the radical 22, cf. Equation (19).
the near future.[31]
By contrast, two equivalents of LiVH2 converts 20 at
However, our studies thus far with certain σ- and π- 25 °C into an 84% yield of 9-fluorenol (23). When such a
bonded organic substrates give insight into the scope and reaction mixture was treated with D2O, the resulting 23a
mode of reaction of 12 and 13, especially when compared was completely deuteriated at C9 and at O. This permits the
2830
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 2825–2835

Epimetallating Reagents for Unsaturated Organic Substrates
activity of benzylic halides toward 12 and 13 and their
pattern of reactions can be readily identified with the inter-
mediacy of radical intermediates. It is not yet clear whether
such radicals are completely “free” of a metal center.
conclusion that the precursor to 23a was the epimetallated
adduct 24, cf. Equation (20). This is an astonishing finding:
despite the presence of V–H bonds, LiVH2 epimetallates,
rather than hydrometallates, 9-fluorenone.
Epoxides
The interaction of 12 or 13 with epoxides, such as styrene
oxide (25) and cis-stilbene oxide (26), again indicate the op-
eration of radical processes, both in the non-stereoselective
deoxygenation of 26 into the thermodynamic 93:7 mixture
of 17 and 16, cf. Equation (21) and in the subsequent poly-
merization of the resultant styrene, cf. Equation (17).
Prospects for the Relative Reducing Actions of VCl and
LiVH2 Compared with those of the Analogous Chromium
Reagents
Halides
Here the interest has been in the reactions of VCl, in
order to note signs of radical involvement. Vanadium(I)
chloride has a low reactivity toward most C–X bonds but
At this stage it can be noted that reducing actions of VCl
can readily insert into benzylic halogen bonds. Thus and LiVH2 appear generally to involve radical intermedi-
2 equivalents of VCl convert benzyl chloride into toluene ates and are distinctly more moderate and usefully selective
(24%) and bibenzyl (76%). Similarly, benzal chloride (27) than those of the chromium reagents. As a telling example,
yields 77% of benzyl chloride and 23% of meso-1,2- the reductive cleavage of σ-C–O, C–S and C–N bonds,
dichloro-1,2-diphenylethane (28). The formation of only the often facile in LiCrH2 reductions, is seldom encountered
meso-isomer may indicate the steric ordering of the chloro- or is much slower with LiVH2. This greater selectivity of
benzyl groups on the vanadium intermediate (possibly 29) reduction with LiVH2 therefore permits the epivanadation
before concerted reductive elimination of the coupled 28 of C=C, C=O and C=N bonds with less undesired σ-bond
takes place, cf. Equation (22). In summary, the superior re- cleavages.
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2831

J. J. Eisch, P. O. Fregene, D. C. Doetschman
FULL PAPER
lated from the neutral combined ether extracts (see above) by ex-
tracting with three 20-mL portions of saturated, aqueous NaHCO3
or 10% aqueous NaOH solution. These extracts were then com-
bined, neutralized with dilute hydrochloric acid, extracted several
times with ether and the combined ether extracts dried with anhy-
drous Na2SO4. The resulting ether extract was then filtered and
freed of solvent to obtain the crude acid product. Product identifi-
cation of known compounds was achieved principally by compari-
son of 1H and 13C NMR spectra of such reaction products with
those spectra of authentic compounds given in the literature.[33,34]
Similarly, product yields from such reactions were calculated based
on integration of the respective NMR proton peaks, except for car-
boxylic acids, where the yields were based on the weight of product
obtained.
Experimental Section
General Experimental Procedures and Starting Materials: All pro-
cedures involving the purification of reaction solvents, distillation
of reagents, the preparations of vanadium(I) chloride (12), lithium
vanadium(I) dihydride (13) and their subsequent measurements or
reactions with organic substrates were conducted under a positive
atmospheric pressure of anhydrous, deoxygenated argon employing
vacuum techniques and with standard Schlenk apparatus. The
Schlenk apparatus typically involved a two- or three-necked Pyrex
flask, with the one neck for the introduction of reaction substrates,
the second for connection with the argon while the third neck could
be coupled to a condenser for heating of reaction mixtures at re-
flux. The drying and deoxygenation of argon, as well as the sol-
vents, such as tetrahydrofuran, toluene and hexane, used in the
reactions were carried out according to established procedures.
Glassware and needles used to transfer liquid samples were dried
in an oven at 140 °C overnight to ensure complete absence of
moisture.[32] Moreover, each needle-equipped syringe was repeat-
edly flushed with argon before every use.
Instrumentation and Analysis: All melting points were measured
with a Thomas–Hoover capillary melting point apparatus and are
uncorrected. Infrared spectra (IR) were recorded with the Perkin–
Elmer spectrophotometers, Model 457 and 283B, which were
equipped with sodium chloride optics. The nuclear magnetic reso-
nance spectra (1H and 13C NMR) were measured in deuteriated
solvents with a Bruker spectrometer, Model AM-360. The 1H
NMR spectroscopic data are reported on the δ scale in parts per
million with reference to an internal standard of tetramethylsilane
(TMS) for solutions in deuteriated chloroform (CDCl3) or deuter-
iated dimethyl sulfoxide ([D6]DMSO) solvent. The 13C NMR spec-
troscopic data are reported on δ scale in parts per million with
reference to deuteriated solvents.
The n-butyllithium in hexane was purchased from Sigma–Aldrich
in Sure SealTM bottles and was used at the stated concentration as
received, uniformly 2.5 . Any punctures made in the rubber sep-
tum for removal of the reagent by gastight syringe were resealed
with wax after each withdrawal. The anhydrous vanadium(III)
chloride was purchased from Sigma–Aldrich at 97% purity and the
vanadium(IV) chloride in at least 99% purity.
Gasometric analysis of samples of VCl/LiCl and LiVH2/LiCl solid
residues were carried out in a Schlenk flask and the evolved gases
were collected in a calibrated gas buret over mercury. The individ-
ual analysis was carried out as follows: Black solutions of either 12
or 13 in THF at room temperature were freed of all volatiles and
THF under an argon atmosphere at reduced pressure until a black
solid remained. Prior to this step, 8 mL of degassed acetic acid was
introduced into the 10-mL Schlenk flask used for protolysis. Flask
A containing solid 12 or 13 and flask B containing degassed acetic
acid were then connected and one outlet joined to the calibrated
gas buret filled with mercury while the other was closed. The pres-
sure in the flask was then balanced against atmospheric pressure
and the initial volume read off to the nearest one-fifth of a milliliter
[mL]. The solid 12 or 13 and acetic acid were allowed to mix, after
which a mildly exothermic gas reaction occurred and the color of
the mixture gradually changed over time from black through olive-
green and finally to green with a white deposit [note: vanadium(II)
chloride is green]. The resulting pressure in the gas buret was then
balanced against the atmospheric pressure, the final volume read
to the nearest one-fifth of a milliliter [mL] and the ambient tem-
perature and atmospheric pressure recorded. Because the concen-
tration of the n-butyllithium was known to two significant figures,
namely 2.5 , the equivalents of dihydrogen evolved upon proto-
lysis could be known only to two significant figures.
Gas chromatographic analyses were performed with a Hewlett–
Packard chromatograph, Model HP5890A, equipped with pro-
grammed temperature rise and with an electronic peak-area inte-
grator and a flame ionization detector (FID).
Electron paramagnetic resonance measurements (EPR) were made
in this department with the use of a CW Varian Spectrometer,
Model V4500, having the X-band set at about 9.2 GHz and fitted
with a Varian Klystron microwave source, model VA297V and a
Varian TE102 cavity.
Mass spectral analyses of the hydrogen gas from the acetolysis of
samples of VCl and LiCrH2 were performed by injecting gas sam-
ples directly into the mass spectrometer of a Hewlett–Packard gas
chromatograph-mass spectrometer, model 5890, series II.
Preparation of Vanadium(I) Chloride (12)
(1) From Vanadium(III) Chloride: In a typical procedure, a purple
suspension of anhydrous VCl3 (330 mg, 2.1 mmol) in 30 mL of an-
hydrous, deoxygenated THF at –78 °C (dry ice/acetone bath) was
treated dropwise with 2.0 equivalents of n-butyllithium in hexane
(2.6 mL of 1.6  solution, 4.2 mmol). The resulting purple mixture
was stirred for 30 min at –78 °C and then brought to room temp.
after which the mixture turned black. After 2 h the VCl (12) in
solution was ready for reaction with individual organic substrates
or the solution could be subjected to high vacuum, in order to
obtain solid 12, admixed with LiCl, which could be subjected to
acetolysis in a gasometric analysis.
The standard hydrolytic workup procedure employed for the usual
reactions of 12 or 13 is as follows. The reaction mixture is treated
with deoxygenated distilled water, 1  or 3  HCl aqueous solutions
or deuterium oxide at room temperature or with methanol at
–78 °C. The organic layer containing the desired products is then
extracted four times with 15 mL of anhydrous diethyl ether. The
combined ether extracts were washed with three 20-mL portions of
saturated, aqueous NaHCO3 or 10% aqueous NaOH solution and
finally with saturated NaCl solution. The combined organic extract
is washed with 20 mL water, dried with solid anhydrous Na2SO4
and then filtered, after which the solvent was removed at the rotary
film evaporator. Carboxylic acids produced in reactions were iso-
(2) From Vanadium(IV) Chloride: In an analogous manner, a red-
dish brown solution of VCl4 (330 mg, 1.7 mmol) in 30 mL of THF
was treated with 3.0 equivalents of n-butyllithium (3.2 mL of 1.6 
hexane solution, 5.1 mmol) at –78 °C, during time the solution
turned from reddish orange to purple. Again, upon warming to
room temp. the solution turned black, consistent with formation
of 12.
2832
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 2825–2835

Epimetallating Reagents for Unsaturated Organic Substrates
Characterization of Vanadium(I) Chloride(12)
By contrast, when gasometric analyses were carried out on solid
samples of 13 prepared from VCl3, the equivalents of H2 gas
evolved ranged between 1.0 and 2.0 of the expected amount of H2.
This results would be consistent with the conclusion that the rea-
gent 13, which has been prepared from VCl3, is much less stable
than the reagent 13 prepared from VCl4. Especially upon solvent
removal the reagent 13 from VCl3 appears to decompose with loss
of H2 to yield lithium hydride, cf. Equations (19) and (20), whose
hydrolysis would provide one equivalent of H2. Any particles of
vanadium metal would react only slowly with glacial acetic acid.
(1) Gasometric Analysis: Treatment of 2.0 to 3.0 mmol samples of
solvent-free solid samples of VCl/LiCl mixtures, prepared either
from VCl3 or VCl4 in the aforementioned manner, with an excess
of glacial acetic acid in a gasometric apparatus (cf. supra) led, from
three samples to the evolution of 0.67, 0.63 and 0.55 mol or an
average 120% of the H2 expected, 0.5 molar equivalent of H2 at
STP according to Equation (9). A typical analysis: VCl prepared
from 2.0 mmol of VCl3 yielded upon acetolysis 34.8 mL of H2 at
297 K and 728 Torr or 0.67 mmol. The mass spectrum of the H2
showed traces of 1-butene (m/z 56), butane (m/z 58) and THF (m/z
72) and indicated the presence of about 90% of H2.
(2) Attempted 51V NMR and EPR Spectra: A solid sample of the
residual mixture of VCl and LiCl was suspended in C6D6. Attempts
to record the 51V NMR spectrum of this sample gave only very
broad absorptions. A similar outcome resulted from attempts to
obtain an EPR spectrum at room temp. However, the presence of
some paramagnetic vanadium component is indicated by such
broad signals.
(2) Infrared Spectroscopic Analysis: The FT-IR spectrum of solid
13 obtained from VCl4, examined in mineral mull under argon,
displayed the following principal peaks (cm–1): ν = 3376 (m), 2921
˜
(m), 2850 (m), 1650 (m), 1511, 1457 (w), 1251 (w), 1000 (m) and
990 (w).
(3) NMR Spectroscopic Analysis: The attempted 51FT-NMR analy-
sis of 13 in C6D6 solution with a VOCl3 reference, generated from
either VCl4 or VCl3, displayed no useful absorption peaks.
(4) EPR Spectroscopy: The reaction product of the butyllithium
reaction with either VCl3 or VCl4, namely LiVH2 (13), was exam-
ined with CW EPR, in tetrahydrofuran and 1:1 toluene-heptane
solutions at room temperature (0.14 ) and with a 1:1 toluene-hep-
tane frozen solution at 140–170 K. Subsequent CW-EPR experi-
ments were done on products in flash frozen: 1:1 toluene/heptane
solvent at 0.14  at 30 K. In addition to the CW-EPR experiments,
a Hahn-echo detected pulsed EPR spectrum was acquired at 30 K.
The experiments were performed with a Bruker ESP 580 X-band
EPR spectrometer equipped with a split-ring ER4AA9 resonator
and were detected with 100 kHz field modulation and phase-sensi-
tive detection. Temperatures below room temperature were
achieved with an Oxford Instruments CF 935 cryostat and an ITC
502 temperature controller. (cf. supra for an analysis of the EPR
measurements.)
(3) Carbonation of Reagent 12: A solution of 6.5 mmol of reagent
VCl (12) in 60 mL of THF at –78 °C was treated with a stream of
dry CO2. Subsequent warming to room temp. and hydrolysis with
5% aqueous NaOH, separation of the aqueous layer, acidifying
with 5% aqueous HCl, and extracting the aqueous layer with ether
gave an ether extract containing 11% of pentanoic acid. This prod-
uct indicates that in the original THF solution VCl has undergone
partial epivanadation with the 1-butene formed by the alkylative
reduction of VCl3, cf. Equation (10).
Preparation of Lithium Vanadium(I) Dihydride (13)
(1) From Vanadium(III) Chloride: In a typical procedure, a purple
suspension of VCl3 (760 mg, 4.8 mmol) in 30 mL of anhydrous,
deoxygenated THF at –78 °C was treated dropwise with 4.0 equiva-
lents of n-butyllithium in hexane (7.7 mL of 1.6  solution,
19.3 mmol). The resulting purple mixture was stirred for 30 min at
–78 °C and then brought to room temp. after which the mixture
turned black. After 2 h the 13 in solution was ready for the re-
cording of its EPR spectrum, for reaction with individual organic
substrates or for the preparation of solid samples of LiVH2, ad-
mixed with LiCl, which could be subjected to acetolysis in a gaso-
metric analysis.
(5) Comparative Reactions of 9-Fuorenone with Lithium Vanadium(I)
Dihydride (13) Prepared from VCl4 or from VCl3: To order to com-
pare the reactivities of LiVH2 (13), prepared either from VCl4 or
from VCl3, toward a typical carbonyl substrate, a standard solution
of 9-fluorenone (20, 650 mg, 3.6 mol) in 10 mL of anhydrous THF
under argon was treated with 3.6 mmol of LiVH2 in 40 mL of THF
at room temp. and stirred for 12 h. In run a the LiVH2 was pre-
pared from VCl4, in run b the LiVH2 originated from VCl3. Typical
hydrolytic workup of each run and addition of diethyl ether allowed
the separation of the organic layer, which after drying with anhy-
drous Na2SO4 and removal of volatiles by rotary evaporation left
(2) From Vanadium(IV) (Chloride): In an analogous manner, a red-
dish brown solution of VCl4, (920 mg, 4.8 mmol) in 30 mL of THF
at –78 °C was treated dropwise with 5.0 equivalents of n-butyllith-
ium in hexane (14.9 mL of 1.6  solution, 23.8 mmol). The re-
sulting mixture was stirred for 30 min at –78 °C and then rapidly
brought to room temp. during which warming the mixture
promptly turned black. After a further stirring of 2 h the reagent
13 was ready for reactions or measurements.
1
a residue that was analyzed by H and 13C NMR spectroscopy.
Results from Run a: Weight of residue: 430 mg, consisting of 37%
of 9-fluorenol, 38% of fluorene, 12% of 9H, 9ЈH-9,9Ј-bifluorenyl,
9% of 9,9Ј-bifluorenylidene and 4% of 9,9Ј-bifluorenyl-9,9Ј-diol.
Results from Run b: Weight of residue: 440 mg, consisting of 57%
of 9-fluorenol, 13% of fluorene, 12% of 9H, 9ЈH-9,9Ј-bifluorenyl,
1% of 9,9Ј-bifluorenylidene and 6% of 9,9Ј-bifluorenyl-9,9Ј-diol.
In addition, 11% of 9-fluorenone was recovered unchanged.
Characterization of Lithium Vanadium(I) Dihydride (13)
(1) Gasometric Analysis: The acetolysis of a solid sample of 13,
prepared from 0.88 mmol of VCl4 and freed from all THF and
other volatiles under reduced pressure, yielded 84.2 mL of H2 at
298 K and under 728 mm mercury. This corresponds to 73.9 mL
or 2.9 mol of H2 at STP. According to Equation (18) the acetolysis
of 13 should afford 2.5 mol of H2. Similar acetolysis provided 2.4,
2.1 and 2.6 mol of H2. (The resulting acetolysis solutions were pale
green.) Mass spectrometric analysis of this H2 showed that it con-
tained traces of THF, butane and 1-butene. The averaged equiva-
lents of H2 from these four runs is 2.5 Ϯ 0.4 equivalents, in reason-
able agreement with the value expected of the empirical formula of
LiVH2 and in light of the aforementioned impurities.
From the recovery of 9-fluorenone and the amount of 9-fluorenol
(57% vs. 37%), it is clear that the LiVH2 prepared from VCl3 (run
b) exhibited a weaker overall reducing action than the LiVH2 pre-
pared from VCl4. This is consistent with other infrared spectro-
scopic and acetolytic gasometric analyses indicating that the LiVH2
derived from VCl3 is thermally unstable even in solution at room
temp.
By contrast, reaction of 9-fluorenone with LiVH2 in a 1:2 molar
ratio in THF at room temp. very similar results regardless of
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2833

J. J. Eisch, P. O. Fregene, D. C. Doetschman
whether LiVH2 from VCl4 or VCl3 is employed. Dimeric reduction drolytic workup gave 290 mg of a mixture of 84% of 9-fluorenol,
FULL PAPER
products were not observed and 83–86% of 9-fluorenol and 12–
15% of fluorene were formed with 1–2% of 9-fluorenone remain-
ing.
(19), 12% of fluorene (20) and 4% of 9-fluorenone.
An identical reaction except that there was a workup with DCl in
97% D2O gave 82% of 9-fluorenol, 14% of fluorene and 4% of 9-
fluorenone. The 9-fluorenol was completely dideuteriated at C9 and
on O.
Chemical Trapping and Characterizing Reactions of Vanadium(I)
Chloride (12)
(1) Isomerization of cis-Stilbene: A solution of VCl (12, 4.2 mmol)
in 30 mL of THF was treated with cis-stilbene (2.52 mg, 1.4 mmol)
at room temp. for 24 h. Hydrolytic workup gave an organic residue
of 90% of trans-stilbene, 4% of cis stilbene and 6% of bibenzyl, cf.
Equation (16).
Acknowledgments
The authors are grateful for the financial support of this research
by the Boulder Scientific Company of Mead, Colorado and by the
Alexander von Humboldt Foundation, Bonn, Germany, which lat-
ter source provided a Senior Scientist Award to J. J. E., enabling
the preparation of this manuscript while the principal investigator
was in residence at the Technische Universität München, Germany
on sabbatical leave. This author remains appreciative of the hospi-
tality extended to him by Professor Wolfgang Herrmann, President
of the TUM, during his stay in the Anorganisch-chemisches Insti-
tut for the academic year of 2005–2006.
(2) Polymerization of Styrene: A solution of VCl (12, 2.8 mmol) in
10 mL of THF was treated with styrene (146 mg, 1.4 mmol) at
room temp. for 12 h. Hydrolytic workup gave only atactic polysty-
rene as a light brown solid, as determined by IR and NMR spec-
1
troscopy, cf. Equation (17). H NMR (CDCl3): δ = 7.30–7.07 (br.
m), 6.8–6.3 (br. m), 2.3–1.7 (m, CH), 1.6–1.1 (m, CH2) ppm. 13C
NMR (CDCl3): δ = 146.0–145.1 (br), 127.9–127.1 (br), 125.8–125.5
(br.), 40.3–40.1 (br.) ppm.
(3) Reduction of 9-Fluorenone (18): A solution of VCl (12,
3.4 mmol) in 40 mL of THF was treated with 9-fluorenone
(300 mg, 1.7 mmol) at room temp. for 12 h. Hydrolytic workup of
one-half of the reaction mixture gave 220 mg of a dark residue solid
that by NMR spectroscopy consisted of 43% of 9,9Ј-bifluorenyl-
9,9Ј-diol, 32% of 9,9Ј-bifluorenylidene, 16% of 9H,9ЈH-9,9Ј-bi-
fluorenyl, 5% of 9-fluorenol and 4% of fluorene, cf. Equation (18).
[1] J. J. Eisch, P. O. Fregene, J. N. Gitua, J. Organomet. Chem.
2007, 692, 4647–4653.
[2] T. Mukaiyama, T. Sato, J. Hanna, Chem. Lett. 1973, 1041.
[3] S. Tyrlik, I. Wolochowicz, Bull. Soc. Chim. Fr. 1973, 2147.
[4] J. E. McMurry, M. P. Fleming, J. Am. Chem. Soc. 1974, 96,
4708.
[5] J. E. McMurry, Acc. Chem. Res. 1983, 16, 405.
[6] The role of low-valent transition-metal salts in organic re-
ductions has received abundant review: a) A. Fürstner, Angew.
Chem. Int. Ed. Engl. 1993, 32, 164; b) A. Fürstner, B. Bog-
danovic, Angew. Chem. Int. Ed. Engl. 1996, 35, 2442; c) A.
Fürstner (Ed.), Active Metals, VCH, Weinheim, 1996.
[7] For a review of the drawbacks and uncertainties inherent in
such empirical procedures for generating subvalent, transition-
metal reductants, consult: J. J. Eisch, X. Shi, J. R. Alila, S. Thi-
ele, Chem. Ber./Recueil 1997, 130, 1175–1187.
Workup of the remaining half of the reaction mixture then showed
that the 9-fluorenol was 95% deuteriated at C9 and on O.
In a reaction run of 2.1 mmol of VCl and 1.7 mmol of 9-fluorenone
in 40 mmol of THF conducted at reflux for 8 h led to 99% of 9,9Ј-
bifluorenylidene, cf. Equation (19).
(4) Benzylic Chlorides: A solution of VCl (6.8 mmol) in 30 mL of
THF was treated with benzyl chloride at room temp. for 12 h. Hy-
drolytic workup and 1H and 13C NMR spectroscopic analysis
showed that 24% of toluene and 76% of bibenzyl had formed.
[8] J. J. Eisch, X. Shi, J. Lasota, Z. Naturforsch., Teil B 1995, 50,
342–350.
[9] J. J. Eisch, S. I. Pombrik, X. Shi, S. C. Wu, Macromol. Symp.
1995, 89, 221.
In a similar reaction between VCl and benzal chloride the reaction
[10] J. J. Eisch, J. N. Gitua, P. O. Otieno, X. Shi, J. Organomet.
Chem. 2001, 624, 229–238.
products were 77% of benzyl chloride and 23% of meso-1,2-
1
dichloro-1,2-diphenylethane, cf. Equation (22). H NMR (CDCl3):
[11] J. J. Eisch, X. Shi, F. A. Owuor, Organometallics 1998, 17,
δ = 7.38–7.24 (m, 10 H) 5.19–5.17 (d, 2 H) ppm. 13C NMR
(CDCl3): δ = 138.3, 128.9, 128.4, 128.0, 65.6 ppm.
5219–5221.
[12] J. J. Eisch, J. R. Alila, Organometallics 1999, 18, 2930–2932.
[13] J. J. Eisch, J. R. Alila, Organometallics 2000, 19, 1211–1231.
[14] O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevski, T. S. Pri-
tyckaja, Zh. Org. Khim. 1989, 25, 2245.
[15] O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevski, Synthesis
1990, 234.
Chemical Trapping and Characterizing Reactions of Lithium Vana-
dium Dihydride (13)
(1) Isomerization of cis-Stilbene: A solution of LiVH2 (3.2 mmol,
from VCl3) in 10 mL of THF was treated with cis-stilbene (576 mg,
3.2 mmol) at room temp. for 12 h. Hydrolytic workup yielded
[16] J. J. Eisch, J. N. Gitua, P. O. Otieno, X. Shi, J. Organomet.
Chem. 2001, 624, 229–238.
1
510 mg of product that by H NMR analysis consisted of 71% of
[17] J. J. Eisch, J. Organomet. Chem. 2001, 617–618, 148–157.
[18] J. J. Eisch, X. Shi, unpublished doctoral studies, 1996.
[19] J. J. Eisch, X. Shi, F. A. Owuor, Organometallics 1998, 17,
5219–5221.
[20] J. J. Eisch, F. A. Owuor, X. Shi, Polyhedron, Symposium-in-
Print 2005, 24, 1325–1339.
[21] For a recent authoritative review on applications of vanadium
salts in organic synthesis, consult: T. Hirao, Chem. Rev. 1997,
97, 2707–2724.
[22] The suggestion by a referee that metallic vanadium could be
present in solid samples of 12 or 13 prior to gasometric proto-
lyses is not supported by the following observations: all such
treatments of samples of 12 or 13 with glacial acetic acid led
rapidly to gas evolution and the formation of a clear green
solution, with no suspended dark particles. In our experience
trans-stilbene and 29% of bibenzyl.
When a solution of LiVH2 (1.9 mmol) in 10 mL of THF was
treated with an excess of cis-stilbene (3.42 g, 19.0 mmol) in a sim-
ilar manner, 3.24 g of product were obtained, which consisted of
99% of trans-stilbene and 1% of bibenzyl.
(2) Polymerization of Styrene: A solution of LiVH2 [4.1 mmol in
40 mL of THF converted styrene (210 mg, 2.0 mmol) into 100% of
atactic polystyrene at 25 °C].
(3) Reactions of 9-Fluorenone with Lithium Vanadium(I) Hydride in
a 1:2 Molar Ratio and Comparative Hydrolytic and Deuteriolytic
Workup: A solution of 3.9 mmol of LiVH2 in 40 mL of THF was
treated with 9-fluorenone (18, 350 mg) at room temp. for 12 h. Hy-
2834
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 2825–2835

Epimetallating Reagents for Unsaturated Organic Substrates
with such protolyses that clearly have particulate metal (Ti, Zr,
Hf, Cr and V), such metal remains unreacted after addition of
HOAc.
[28] A. Abragam, B. Bleaney, Electron Paramagnetic Resonance of
Transition Metal Ions, Dover, New York, 1986.
[29] Other coordination configurations about VI would also accom-
modate a linear H–V–H bonding: 1) d sp2-square planar; and
2) d sp3-trigonal bipyramidal. But the higher coordination of
d2sp3 (octahedral) seems to be favorable here.
[23] Although the by-product LiCl through its Cl– and the solvent
THF may well be coordinated with 12 and/or 13 in solution,
the gasometric protolysis of solid samples of 12 or 13 demon-
strates that such solvated THF can be removed in high vac-
uum. As to chemical complexation of LiCl with samples of 12
and 13 in the solid state, such stable complexes are unlikely. In
previous work, samples of TiCl2·2THF, ZrCl2·2THF
andLiCrH2·2THF, free of LiCl, have been isolated from the
solvent pair of THF and toluene, cf. ref.[8,13,19]
[30] S. P. McGlynn, T. Azumi, M. Kinoshita, Molecular Spec-
troscopy of the Triple State, Prentice-Hall, Englewood Cliffs,
NJ, 1969.
[31] J. J. Eisch and P. O. Fregene, manuscript in preparation, al-
though such results concerning the epivanadation of C=O and
C=N linkages by LiVH2 have already been adumbrated in
ref.[1]
[24] C. J. Curtis, J. C. Smart, J. L. Robbins, Organometallics 1985,
4, 1283.
[32] Detailed descriptions of appropriate techniques for working
with air- and moisture-sensitive organometallics can be found
in: J. J. Eisch, Organometallic Syntheses, vol. 2, Academic Press,
New York, 1981, pp. 1–84.
[25] G. Wilkinson, F. G. A. Stone, E. W. Abels (Eds.), Comprehen-
sive Organometallic Chemistry I, vol. 3, Pergamon, New York,
1982, pp. 647.
[26] Consult the reference J. J. Eisch, J. N. Gitua, D. C. Doetsch-
man, Eur. J. Inorg. Chem. 2006, 1968–1975, for the EPR studies
on the open-chain biradical of the titanium(II) isopropoxide
trimer and the probability that diamagnetic titanium(II) chlo-
ride also exists as a trimer in a corresponding three-membered
ring.
1
[33] J. C. Pourhert, J. Behnke (Eds.), Aldrich Library of 13C and H
FT-NMR Spectra, vol. 1–2, first edition, Aldrich Chemical Co.,
Milwaukee, 1993.
1
[34] J. C. Pourhert, J. Behnke (Eds.), Aldrich Library of 13C and H
FT-NMR Spectra, third edition, Aldrich Chemical Co., Mil-
waukee, 1981.
[27] A. Carrington, A. D. McLachlan, Introduction to Magnetic
Resonance, Wiley, New York, 1979.
Received: January 28, 2008
Published Online: April 17, 2008
Eur. J. Org. Chem. 2008, 2825–2835
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2835